Hello Paul.
An upload to Experimental of MLMMJ "1.5.0-1+really" which disables the
upstream post-build tests seems like a good idea, as it seems those
tests are unreliable. I want to "convert" at least some of those tests
into autopkgtests eventually. Hopefully those tests won't show the same
unreliability that the atf-sh tests do. Neither I nor upstream
understand what the issue is with buildd results from the atf-sh tests.
It doesn't make sense that those seem to work "sometimes".
As far as I'm aware the PHP code for mlmmj-php-web and
mlmmj-php-web-admin in MLMMJ 1.5 has been updated, so those packages
shouldn't need removal and don't need bug reports. The issue is with the
old PHP code in MLMMJ 1.3.0-4.
I don't have the information I would need at the moment to open bugs on
mlmmj-php-web and mlmmj-php-web-admin in 1.3.0-4. First I'm asking the
person that reported them to be broken to open a bug report, and I won't
count on that happening. I'm planning on building an i386 VM to test
that MLMMJ 1.5.0 functions as expected in order to verify that the build
test failures are likely false positives; I'll also see if I can verify
the PHP failures in mlmmj 1.3.0-4 while I'm doing that.
That's what I can do.
-- Chris
Chris Knadle
chris.kna...@coredump.us
On 5/24/25 05:48, Paul Gevers wrote:
Hi Chris,
On 22-05-2025 16:58, Chris Knadle wrote:
Other than that as far as I'm aware the mailing list code still
works. The configuration for MLMMJ is via text files, so it's
possible to configure it without using mlmmj- php-web and
mlmmj-php-web-admin.
Maybe not what you were looking for, but maybe it's the best course of
action *now* to do your proposed upload to experimental [1, 2] and to
upload a +really version of the the package currently in testing
dropping the problematic binaries? As you hinted at and judging from
popcon, most users don't have the php packages installed.
Either way, it would be good to document the issues against the
binaries in the bts.
Paul
[1] to prevent a round trip through NEW once you add them back in
unstable
[2] with a higher version than the +really version for unstable
